Presentation is loading. Please wait.

Presentation is loading. Please wait.

Boise N.F. Lessons Learned John Mootz APFO 2008 USDA Imagery Planning Meeting December 3, 2008.

Similar presentations


Presentation on theme: "Boise N.F. Lessons Learned John Mootz APFO 2008 USDA Imagery Planning Meeting December 3, 2008."— Presentation transcript:

1 Boise N.F. Lessons Learned John Mootz APFO 2008 USDA Imagery Planning Meeting December 3, 2008

2 12/3/20082008 USDA Imagery Planning Meeting2 Topics  Flight Planning  Image File Naming  Photo Center File  Color Bit Depth  Color Band Order  Image Rotation  Inertial Measurement Unit  Camera Calibration Report

3 12/3/20082008 USDA Imagery Planning Meeting3 Topics not discussed  Delivery media  Provisioning/distribution  Long-term storage

4 12/3/20082008 USDA Imagery Planning Meeting4 Flight Planning Background  Traditional film acquisition had well established flight planning process 9-inch squared aerial film North-south flights Footprint was based on film scale Endlap: 62%; Sidelap: 30%  APFO flight planned and provided exposure stations with contract

5 12/3/20082008 USDA Imagery Planning Meeting5  Government cannot “flight plan” Each camera has unique footprint Camera not known until after award  Flight planning requirement passed to contractor Allows coverage optimization (based on aircraft, terrain, weather patterns, etc) Example: 2008 NAIP TN Flight Planning (Con’t) Direct-digital Cameras

6 12/3/20082008 USDA Imagery Planning Meeting6 Flight Planning (Con’t) Direct-digital Camera Footprints DMC 7,680 x 13,824 120mm/25mm UltraCamXp 11,310 x 17,310 100mm/33mm ADS-80 12,000 line 64° FOV

7 12/3/20082008 USDA Imagery Planning Meeting7 Flight Planning (Con’t) Lesson Learned  Boise project required contractor to submit flight plan for approval Due to the extreme terrain, originally planned for 80%/56% (block D) Reduced to 72%/43% Extra coverage increased acquisition risk and file sizes  Sidelap/Endlap significantly impacts file sizes

8 12/3/20082008 USDA Imagery Planning Meeting8 Image File Naming Background  Original naming convention _ _.tif project code = project code image number = consecutively numbered value yymmdd = image exposure date Example:614020_00001_20080827.tif  Mod 2 naming convention _ _ _.tif project code = project code flt line no = flight line number exp no = consecutively numbered value yyyymmdd = image exposure date Example:614020_0025_0001_20080827.tif

9 12/3/20082008 USDA Imagery Planning Meeting9 Image File Naming (Con’t) Lesson Learned  Predetermined stations not possible Each direct-digital system has different footprint  Need a way to index back to location  Consistency is needed within project Photo Center File numbers must match  Only one file allowed per “station”

10 12/3/20082008 USDA Imagery Planning Meeting10 Photo Center File Film-based Acquisition Project Identification Code6 Flight Line Number*4 Exposure Number*4 Date of Exposure8 Time of Exposure – Local 24 Hour Clock6 Sensor Serial Number **15 Latitude (DD.DDDDD)8 Longitude (-DDD.DDDDD (Negative))10 Flight Altitude in meters at camera8 *Same flight line and image numbers used for file naming convention.. MAXIMUM NUMBER OF DESCRIPTIONCHARACTERS IN FIELD ±30m accuracy

11 12/3/20082008 USDA Imagery Planning Meeting11 Photo Center File (Con’t) Direct-Digital Project Identification Code6 Flight Line Number*4 Exposure Number*4 Date of Exposure8 Time of Exposure – Local 24 Hour Clock6 Sensor Serial Number **15 Latitude (DD.DDDDD)8 Longitude (-DDD.DDDDD (Negative))10 Flight Altitude in meters at camera8 Number of GPS Satellites Acquired2 Position Dilution of Precision (PDOP)3 IMU omega value (Radians)10 IMU phi value (Radians)10 IMU kappa value (Radians)10 *Same flight line and image numbers used for file naming convention. **If digital camera has more than one sensor head please use the camera serial number. Example: 614020,25,1,20080827,130755,12345678,42.71936,-23.41498,7048.63,5,1.5,.0001358,.01073000,-.8732658 MAXIMUM NUMBER OF DESCRIPTIONCHARACTERS IN FIELD ±5m accuracy

12 12/3/20082008 USDA Imagery Planning Meeting12 Photo Center File (Con’t) Lesson Learned  PCF is critical for spatial positioning  GPS/IMU data needed  5-decimal place lat/lon is only accuracy to approximately 1m

13 12/3/20082008 USDA Imagery Planning Meeting13 Color Bit Depth Background  Most digital cameras acquire imagery at 12 or 14-bits per color  Graphic file formats use bytes (8- bits) to store each pixel color  Bit depth determines the number of possible “shades” per color 256 shades for 8-bit (2^8) 65,536 shades for 16-bit (2^16)

14 12/3/20082008 USDA Imagery Planning Meeting14 Color Bit Depth (Con’t) GIS vs Remote Sensing  GIS users: Want “pretty picture” 8-bit, color balanced image  Remote Sensing users: Want unmodified radiometric data No data loss 16-bit, unstretched image  Contradictory needs

15 12/3/20082008 USDA Imagery Planning Meeting15 Color Bit Depth (Con’t) Histogram Basics Highlights Shadows Midtones A graph showing the relative distribution of various tonal values in an image. The histogram shows the number of “dark” values on the left and “bright” values on the right. Its purpose is to show the distribution of tone throughout an image.

16 12/3/20082008 USDA Imagery Planning Meeting16 Color Bit Depth (Con’t) Histogram Examples DARK IMAGE FLAT IMAGE LIGHT IMAGE Shifted Compressed

17 12/3/20082008 USDA Imagery Planning Meeting17 Color Bit Depth (Con’t) 12-bit Storage in 16-bit File Format 12-bit: 02,0484,096 16-bit: 032,76865,535 016,38432,76849,15265,536 Approx 6%

18 12/3/20082008 USDA Imagery Planning Meeting18 Color Bit Depth (Con’t) Unstretched Image Unadjusted Boise NF image displayed in ArcMap

19 12/3/20082008 USDA Imagery Planning Meeting19 Color Bit Depth (Con’t) Histogram Stretching Stretch (LUT)

20 12/3/20082008 USDA Imagery Planning Meeting20 Color Bit Depth (Con’t) Histogram Comparison

21 12/3/20082008 USDA Imagery Planning Meeting21 Color Bit Depth (Con’t) Image Options  Do nothing ArcMap does not automatically apply histogram stretch  Pre-stretch histogram Stretching is irreversible Data loss may impact RS users  Create an “AUX” file Contains statistic information

22 12/3/20082008 USDA Imagery Planning Meeting22 Color Bit Depth (Con’t) Histogram Stretching LinearNon-linear

23 12/3/20082008 USDA Imagery Planning Meeting23 Color Bit Depth (Con’t) Lesson Learned  Unprocessed imaging will look “black” without stretching  AUX file will resolve issue in ArcMap  More than one method of “stretching” histogram  Users need to be educated

24 12/3/20082008 USDA Imagery Planning Meeting24 Image Rotation Sensor Attitude  Long dimension of image is flown parallel with a/c wings  Reduces number of flight lines (lowers acq cost and risk)  “Upper left” corner of image may not be the northwest point Source: Aero-Metric

25 12/3/20082008 USDA Imagery Planning Meeting25 Image Rotation(Con’t) North-South Flight Lines N Northwest corner North bound N Northwest corner South bound A/C lower left N N DMC saves images in “portrait” mode

26 12/3/20082008 USDA Imagery Planning Meeting26 Image Rotation(Con’t) Importing Unprocessed Imagery  RS software will not “auto” rotate image using the GeoTIFF/TFW rotation information N ArcMapSocet Set N North bound

27 12/3/20082008 USDA Imagery Planning Meeting27 Image Rotation(Con’t) Lesson Learned  RS software requires applying the GPS/IMU data to display north up  Cannot “flip” image because correct image/IMU orientation would be lost  End users will have to understand that each camera manufacturer may save image data differently

28 12/3/20082008 USDA Imagery Planning Meeting28 Color Band Order Displaying Multispectral Imagery  Remote sensing software assumes ”satellite” band order when opening multispectral imagery ArcMap 1234 123 ERDAS 4-band NAIP imagery 432

29 12/3/20082008 USDA Imagery Planning Meeting29 Color Band Order (Con’t) Lesson Learned  No issue with 3-band imagery  GIS software applications act consistently with photographic applications  One set of users will have to adapt RS users are more “experienced” with band order  Long-term consistency is important

30 12/3/20082008 USDA Imagery Planning Meeting30 Inertial Measurement Unit General Concept  An IMU works by detecting the current rate of acceleration, as well as changes in rotational attributes, including pitch, roll and yaw. This data is then fed into a computer, which calculates the current speed and position, given a known initial speed and position  A major disadvantage of IMUs is that they typically suffer from accumulated error. Source: www.wikipedia.com

31 12/3/20082008 USDA Imagery Planning Meeting31 Inertial Measurement Unit (Con’t) IMU Errors Dead Reckoning Est Final Error Reverse filtering Kalman Filter

32 12/3/20082008 USDA Imagery Planning Meeting32 Inertial Measurement Unit (Con’t) Differential GPS vs IMU AdvantageDisadvantages DGPS High accuracy of position and velocity estimation Time-independent error model Low bandwidth Satellite shading (dropouts) Slow ambiguity resolution IMU Full 6 DOF solution Continuous data acquisition Self-contained (no dropouts) Drift - solution errors grow over time INS / DGPS Combine all advantages of both systems Redundant and complementary data (both systems’ errors are separately observable) Navigation through GPS outages GPS fixes allow INS error estimation No significant limitations

33 12/3/20082008 USDA Imagery Planning Meeting33  Inertial navigation system starts from a known position (i.e. airport)  Current position is updated by IMU  Drift error is corrected with GPS  Post-processed for backward solution Inertial Measurement Unit (Con’t) Combining IMU & ABGPS Source: Applanix Corp

34 12/3/20082008 USDA Imagery Planning Meeting34 Inertial Measurement Unit (Con’t) Lesson Learned  Must rely on contractor’s to accurately post-process data No method for inspecting accuracy CORS/Base stations impact accuracy  Processed GPS/IMU data very important for ortho production  No industry standard for text data Contract specification will have to state what software end user is using

35 12/3/20082008 USDA Imagery Planning Meeting35 Camera Calibration Report Lesson Learned  No report similar to film-based USGS camera calibration report  Must rely on manufacturer’s calibration report  No “fiducial marks”  Frame-based cameras have multiple sensors DMC has 8 sensors/2 lens lengths

36 12/3/20082008 USDA Imagery Planning Meeting36 Summary  Intended use of imagery significantly drive requirements  Need to standardize products for consistency  End users need basic level of education

37 12/3/20082008 USDA Imagery Planning Meeting37 Questions


Download ppt "Boise N.F. Lessons Learned John Mootz APFO 2008 USDA Imagery Planning Meeting December 3, 2008."

Similar presentations


Ads by Google